Molecular Rearrangements. VIII. a Mechanistic Correlation of the

Molecular Rearrangements. VIII. a Mechanistic Correlation of the Aldehyde-Ketone and Pinacol Rearrangements1. Ben M. Benjamin, Clair J. Collins...
0 downloads 0 Views 1MB Size
4329

-~LDEHYDE-KETONE A N D PINACOL REARRANGEMENTS

Sept. 5, 1956

[CONTRIBUTION FROM

THE

CHEMISTRY DIVISIONO F O A K RIDGEXATIONAL LABORATORY]

Molecular Rearrangements. VIII. A Mechanistic Correlation of the Aldehyde-Ketone and Pinacol Rearrangements1 BY BENM. BENJAMIN AND CLAIRJ. COLLINS RECEIVED FEBRUARY 16, 1956 Diphenyl-p-tolylacetaldehyde-l-Cl4 (IY) , chain- and ring-labeled 1,2-diphenyl-l-p-tolylethylene glycol (Va, Vb and VC) and l,l-diphenyl-2-p-tolylethylene-2-C14 glycol (VIa) have been prepared and subjected t o rearrangement (a) in cold, concentrated sulfuric acid and (b) in boiling formic acid. Under both sets of conditions two ketones, p-tolyldesoxybenzoin (YII) and benzhydryl p-tolyl ketone (VIII), are produced from 11-, \- or 1-1. By using double-labeling techniques and by generating the same intermediate ion (XI, Charts I, I1 and 111) under identical conditions from the three different sources (IY,Y and \'I), it has been possible t o relate the pinacol and aldehyde-ketone rearrangements mechanistically. For the rearrangement of the aldehyde IY, the simple ratios of the products YII/VIII (4.7 in sulfuric acid and 3.5 in formic acid) would appear to indicate phenyllp-tolyl migration ratios of 2.35 and 1.75, respectively, which are inverted with respect to the usual migratory abilities of the two groups, It has been shown t h a t the migration ratio of the one p-tolyl and the two phenyl groups of aldehyde 11-is not reversed and is not given by the ratio of the ketonic products. The true p-tolyl/phenyl migration abilities in the rearrangement of 11. in either sulfuric or formic acid have been estimated by the use of equation 8 to be greater than one as normally expected. It has been shown t h a t secondary hydroxyl removal from 1,l-diphenyl-Zp-tolylethylene-2-C14 glycol (1.1a) is relatively more important, with respect to tertiary hydroxyl removal, in concentrated sulfuric acid at 0". than in formic acid a t 27.6".

Introduction The acid-catalyzed rearrangements of di- and trisubstituted acetaldehydes to produce ketones have long been regarded as intimately related to the pinacol rearrangement,2-5 since under relatively mild conditions aldehydes are often products of the pinacol transformation, whereas under more strongly acidic conditions, the appropriate aldehyde is capable of conversion to the same ketone or ketones obtained from the parent pinacol. Neither the pinacol nor the aldehyde-ketone rearrangement is well understood. The difficulties involved in understanding the former have been most adequately presented by U'heland. The anomalies relative to the aldehyde-ketone reaction have been mentioned both by Wheland4 and by Ingoldj and have to do with the fact that in this rearrangement there is an apparent reversal of the usual migratory abilities of the groups i n ~ o l v e d . ~ JThus, di(Me)2CCHO

I

Ph

MeCHCOMe

I

Ph

OH OH

I

OH OH

*I

Ph2C6HO I

PhC-CHPh I

CHI

CH3 Va

PhL*-CHPh

Me2CHCOPh

I1

I11

methylphenylacetaldehyde (I) in concentrated sulfuric acid yields 3-phenyl-2-butanone (11) rather than phenyl isopropyl ketone (111)6despite the well-known' fact that phenyl ordinarily possesses a migratory ability much greater than methyl. Nor is this seeming reversal of the migratory abilities confined to aldehyde I, for the phenomenon is appar(1) This paper is based upon work performed under Contract Number W-7405-eng-26 for t h e Atomic Energy Commission a t Oak Ridge National Laboratory. Paper V I I , J. F. Eastham, J. F. Huffaker, V. F. Raaen and C. J. Collins, THISJ O U R N A L , 78, 4323 (1956). ( 2 ) (a) S.Danilov, J . Rziss. Phys. Chem. Soc., 49, 282 (1917); 61, 97 (1919); (b) S.Danilov and E. Venus-Danilova, Bey., SSB, 377-387 (1926). (3) R Roger and W. B. RlcKay, J . Chem. SOC.,332 (1933). (4) G. W. Wheland, "Advanced Organic Chemistry," 2d Ed., John Wiley and S o n s , Inc., h'ew York, N. Y., 1949. pp. 494-534. ( 5 ) C. K. Ingold. ''Structure and Mechanism in Organic Chemistry," Cornel1 University Press, Ithaca, K.Y . , 1953, p. 479. (6) A. Orekhov and M. Tiffeneau, C o m p t . r e n d . , 182, 67 (1926). (7) See, for example, hl. Tiffeneau and H. Cahmann, Bull. SOC. chim., 1876 (1935), and t h e discussion of D. Y . Curtin and S.Schmukler, THISJ O U R N A L , 77, 1107 (195.5), who state "that 10 is a conservative estimate for t h e phenyl'alkyl migration ratio , , in t h e deamination . I '

of l - a m i n o ~ 2 - p h r n y l p r o p a n o i 2

ently quite generaL8 In a recent paperg the importance of the conjugate acid of triphenylacetaldehyde as an intermediate in the pinacol rearrangement of triphenylethylene glycol was evaluated for several different acid catalysts. It appeared to us that these resultsg foreshadowed a general mechanism which would allow a correlation between the pinacol and aldehyde-ketone rearrangements and which would further clear up some of the mystery4J which has been associated with these two reactions. Accordingly, we have applied the same type of radiochemical investigationgto the acid-catalyzed rearrangements of diphenyl-p-tolylacetaldehyde-lC14 (IV) and the closely related glycols V I O and These compounds were selected as models for this study, since the migratory abilities of p-

rv

OH

* I

OH

I

CHI \'b OH

OH

PhC-CHPh

I

VIa

(8) For other examples of t h e aldehyde-ketone rearrangement in which t h e usual migratory abilities416 of t h e groups are reversed, see: S. N. Danilov, J . Rcss. Phys. Chem. Soc., 61, 109 (1919); 67, 347 (1925); 68,120 (1926); 69, 196, 210 (1927); 61, 723 (1929); J. Gcn. Chcm., 18, 200 (1948); R . Lucas and J. Guerlain, Bull. soc. chim.,49, 1860 (1931): and K. Isimura, Bull. chem. snc J a p a n , 16, 196, 252 (1941); M . Tiffeneau and A. OrBkhoff. Compi r e n d . , 182, 67 (1926). (9) Paper V, C. J. Collin's, THISJ O U R N A L , 77, 5517 (1955). ( 1 0 ) (a) S.F. Acree, B e r . , 8 7 , 2753 (1904); ( b ) A. McKenzie, A. K. Mills and J . R . Myles, t b i d . , 6 3 , 904 (1930); (c) C. F. Koelsch. THIS J O U R N A L , 64, 2049 (1932); (d) A. McKenzie, R . Roger and W. B. McKay, J . Chem. Sot., 2597 (1932). (11) T h e asterisks denote t h e positions of carbon-14 labeling, T h e lower-case letters after t h e Roman numerals indicate specific isotopeposition isomers. I n t h e case of compound V , \'a was the p-diastereomer, whereas Vb and Vc mere t h e 0-diastereomers

BEN hl. BENJAMIN AND CLAIR J. COLLINS

4330

tolyl and phenyl do not differ so greatly1*as those of the substituents of aldehyde I, thus permitting an accurate experimental determination of the yields of ketone VII, formed from IV by phenyl migration, and ketone VIII, formed by p-tolyl migration.

CH, k711

VI11

When diphenyl- p - tolylacetaldehyde (IV), labeled with the carbon-14 in the carbonyl group, was treated with cold, concentrated sulfuric acid, the mixture of ketones VI1 and VIII, as determined by the radioactivity dilution method, l 3 was obtained in the ratio VII/VIII = 4.7. m'hen the aldehyde IV was treated with boiling formic acid, the same two ketones were obtained quantitatively in the ratio VII/VIII = 3.5. I n each case that ketone (VII) which was formed by phenyl migration predominated over the ketone (VIII) which was formed with p-tolyl migration. Allowing for the statistical factor of two phenyls and only one p-tolyl group in aldehyde IV, in the sulfuric acid experiment the apparent phenyllp-tolyl migration ratio is 2.35, whereas in the formic acid experiment this ratio is 1.73. I t is difficult to believe that these two ratios represent the true p-tolyllphenyl migratory abilities when the mass of data previously citedI2 is considered. Results obtainedg for the rearrangement of triphenylethylene glycol seemed to indicate that the rearrangement of aldehyde I V to ketones VI1 and VI11 could best be explained by the following mechanism, in which X1, Xz and Xa represent three ionic species in a dynamic equilibrium : OH PhZCCHO--,

CH3

@

kP

o x ,

k+

Ph2CCHOH

CH3

81

PhCCHPh

kH

+ PhCHCOPh

Qm

Qx2 CH3

CH3

k iJlkmL

x3

Vol.

is

jected to acid catalysis. It is further apparent, even though the ratio 2 k ~ I k pmight be expected to be greater t h a n one, that this ratio alone is not sufficient to control the relative yields of the ketones VI1 and VIII, since the additional rate constants k+, kTol, k H and k ' H would also be expected to affect the final yields of products. The validity of the foregoing mechanism, together with its relation to the pinacol rearrangement, is established in the folIowing sections. Methods and Results The aldehyde I V and glycols Va, Vb and VIa were subjected to rearrangement (a) in cold, concentrated sulfuric acid and (b) in boiling formic acid. The yields of ketones VI1 and VI11 were determined by the radioactivity dilution method.I3 The results of these experiments are given in Table I. Compounds IV, Va and VIa were subjected to TABLE I 1'111 PRODUCED B Y THE l\CTION FORMIC OR SULFURIC ACID UPON lY,V ASD VI

YIELDS OF KETONES 1.11 A S D OF

No. 1 2

3 4 5 6 7

8 Q 10

Reactant

iv

Yield of ketone,a % VI1 VI11 17.3 82.5 22.3 77.7 86 14 16.4 83.6 85.1 14.9

Reagent

Ratio VII/ VI11 4.7 3.48 6.15 5.11 5.82

IV Va, p form Va, &form Vb, a-form

96% HzSO, a t 0' Boiling 98% formic 96% His01 a t 0' 96% HzSOj a t 0' 96% HzSOi at 0'

Va, @-form Vb, a-form

Av. Boiling 987' formic Boiling 98% formic

84.9 87.7 88.3 __

15.1 12.3 11.7

_-

7.13 7.57

VIa VIa VIa

Av. 9617, H z S O i a t O O Boiling 98Y0 formic Boiling 98% formic

88.0 83.0 71.7 71 . O

12.0 17.0 28.3 20.0

4.9 2.54 2.49

_

Av.

_

__

--

71.4

28.6

~

The values give11 in these two columns represent the j-iclds of ketones in the products, and thus for each experiment the yields of 1'11 and VI11 will total 100. I n the sulfuric acid-catalyzed reactions the combined yields of products were 94-99?;,. The combined yields in the formic acidcatalyzed reactions were quantitative.

.the same conditions in different experiments, and in each experiment the mixed ketones were cleaved with alkali

JllIT

It is apparent that if this scheme i s correct, the specific rate-constant ratio kT/kp, rather than the product ratio VIII/VII, represents the relative migratory abilities of the one p-tolyl and the two phenyl groups of aldehyde IV when this aldehyde is sub(12) J. G. Burr, THIS J O U R N A L , 7 6 , 5008 (1953j, has shown t h e P tolyl/phenyl migration ratio to be 2.5 in t h e acetolysis of 2-phenyl-2~ - t o l y l e t h y l - l - C ltosylate. ~ J . G. Burr and L. S. Ciereszko, ibid., 74, 5426 (19521, have determined a ratio of 2.0 in t h e dehydration of 2-phenyl-2-p-tolylethanol-l-C~*. For t h e Schmidt reaction t h e p tolyl/phenyl migration ratios have been variously reported as 3.44 [ R . F. Tietz and W. F. McEwen, i b i d . , 4011 (1955j1: 5.0 [W. E. M c Ewen, M. Gilliland and B. I. Sparr, i b i d . , 72, 3212 (1950)l and 4 0 IS. N. Ege and K . W. Sherk, ibid., 76, 354 (1953)l. I n t h e deaminaD. Y . Curtin and M. C. tion of 2-amino-l-phenyl-l-~.tolylethanol, Crew, i b i d . , 76, 3719 (1954), have reported a .i-tolyl/phenyl migration ratio of 1.3. Thus, a good average value for this migration ratio would seem t o be 3.0. (13) R . H Rlayor and C. J . Collins, i b i d , 73, 471 (1951).

IXa

+ PhEOOH

CH3 YIIab

CHR 1-1IIa b

Xa

The neutral fraction, consisting of phenyl-fi-tolylmethane (IXa) and diphenylmethane (Xa), was then oxidized with chromic acid

Sept. 5 , 1956

433 1

ALDEHYDE-KETONE AND PINACOL REARRANGEMENTS

* COPh I

In the latter instance the neutral (benzophenone) fractions were discarded, and only the ring-labeled p-benzoylbenzoic acid fractions were purified and assayed. The results of these experiments are given also in Table 11. Several additional experiments were then performed upon the glycols V and VI using as a solvent 98yo formic acid a t 27.6 &OOH 0.1'. Under these conditions the aldehyde IV is a The benzophenone, separated as the neutral frac- product from each glycol and is quite stable. The tion, was assayed for its radioactivity content aldehyde IV produced in each experiment could be whenever possible as the 2,4-dinitrophenylhydra- determined by conversion nearly quantitatively to zone. This fraction was always difficult to isolate, its 2,4-dinitrophenylhydrazone under conditions since it was the minor constituent of the reaction which did not favor the formation of the derivaproduct. The p-benzoylbenzoic acid was carefully tives of VI1 and VIII. The yields of VI1 and VI11 purified by repeated crystallizations and then as- were determined by means of the radioactivity dilusayed for its radioactivity content. The results of tion techniques.13 The results of these experithese experiments are given in Table 11. ments are given in Table 111. TABLE 11 TABLE I11 MOLARRADIOACTIVITIES OF THE OXIDATION FRAGMENTS OF IV OR KETONESVI1 AND KETONES VI1 AND VI11 OBTAINED B Y THE A C T I O N O F H Z S O ~ P E R CENT. YIELDS O F ALDEHYDE ACIDAT 27.6' UPONV OR \'I VI11 FROM ACTIONOF FORMIC OR HCOOH UPONIV, V, VI OR VI1

No.

Reactant"

Acid reagent

Activity of benzoyl- Activity of benzoic benzoacid phenone from VI1 from VI11

IV 96% H?SO4 a t 0" 0.013 .. . IV Boiling formic .115 ... Va, @-form 96",, HzSO4 a t 0' ,018 0.015 Vb, a-form Boiling formic ,955 ,960 5 Va, &form Boiling formic .040 .038 6 VIa 96% HzSO4 a t 0" ,278 ... 7 VIa Boiling formic .174 . . 8 VIa Boiling formic .165 ... 9 VIa Boiling formic .165 .. . 10 Vc, P-form 96% H*SOI a t 0" .563 ... 11 Vc, &form Boiling formic .765 ... 12 VIIdb 96% H2S04 a t 0' 0 ... The molar radioactivities of the reactants are taken as 1.OOO mc. per mole; the molar radioactivities of the degradation products from each reaction are taken as fractions thereof. * p-Tolyldesoxybenzoin labeled solely in the phenyl group adjacent to the carbonyl. 1 2 3 4

Glycol Vc was next accorded the same treatment just described for the chain-labeled glycol, and the ketonic products were subjected to the identical degradative scheme. * hhCHCOPc I

PhCHz I

Yield

Compound

Va Va Vb Vb Va T'a

VIa VIa VIa VIa

I

I

CHI I X C COOH

VIIIcd

XC

I

CHa COPh*

I

COOH

deriv.

Reagent

Formic Formic Formic Formic Formic

acid, acid, acid, acid, acid,

27.6' 27.6' 27.6" 27.6' 27.6' Av. Formic acid, 27.6' for 24 hr., then boiled 6 hours Formic acid, 27.6' Formic acid, 27.6' Formic acid, 27.6' Formic acid, 27.6'

of IV, %

Cor-

rected yield, %of aldehyde

Yield, %

VI1

VI11

51 .O 44.4 44.9 43.6 44.9 49"'a 44.5

88.3 11.7 5. 05d 83.4 83.7 84.7

-

87"9c

Av. 83.9 Although the pure aldehyde could be converted quantitatively to its 2,4-dinitrophenylhydrazone,mixtures of IV, VI1 and VIII did not give this derivative quantitatively, and the smaller the proportion of IV, the lower the yield. The yield of aldehyde has been taken as 100 - 51 = 49Y0. The yield of 2,4-dinitrophenylhydrazonehas been divided by the correction factor 0.975. Oxidation of this sample followed by radioactivity assay of the p-benzoylbenzoic acid obtained demonstrated that 100% of the molar radioactivity of the ketone VIIb was present in the 4-methylbenzhydryl portion of the molecule.

Discussion The Sulfuric Acid-catalyzed Reaction.Following are the implications of the pertinent data of the preceding section: 1. Inspection of Tables I and I1 reveals the yield of each ketone from the different reactants, as well as the fates of the various carbon-14 labels. Given in Chart I is a mechanism for the rearrangements of IV and V which accounts for these facts and which is similar to that proposedgfor the reaction of triphenylethylene glycol itself in cold, concentrated sulfuric acid. The consequences both of phenyl and of chain labeling are indicated. 2 . The contribution of each reaction path of Chart I can now be calculated (see Experimental section). If ma, m b , mc, m d and me are the moles of glycol V which proceed to VI1 or VI11 through

A.

CH3 VIIcd

of 2,4DNPH

BEN 11.BENJAMIN AND CLAIR J. COLLINS

4332

Vol. i S

paths A, B, C, D and E, then m a = 0.017, mb = Thus 0.003; m, = 0.124, VZd = 0.708, “2, = 0.148. If our mechanism is correct, the ratio m d / m e should be the same, within experimental error, as the ratio of VII/VIII obtained from the aldehyde IV under the same conditions. From the foregoing results we see that m d / m e is 4.8, in excellent agreement with At complete reaction the ratio m d ’ m e is given by the value 4.7 (Table I, line 1) for the ketonic ratio E d - kaSs (7) obtained from aldehyde IV. ffl, k’HS3 3 . From Chart I it seems likely that ion Xz, stabilized by phenyl and p-tolyl groups, should have I t then readily follows14that a higher stability than ion X3, stabilized by two CHARTI

A

B

kH

PhCCHPh

OH I

@I*

Ph$ CH

PhCHtOPh

The ratio k T o l / k ’ H may be evaluated in principle by considering the rearrangement of glycol VIa in cold, concentrated sulfuric acid, in a manner which will be discussed later in this section for the formic acid-catalyzed reactions. The ratio kTol ’k’H for the sulfuric acid-catalyzed reaction, however, is SO large that the error inherent in the calculation precludes an accurate experimental determination. I t is possible, despite this fact, to estimate probable values for kTol k ’ H from available data. Given in Table IV are k a r / k H ratios for three different carbonium ions. Itre assume that in the sulfuric acid-cata-

0 -k CH3

*Ph2CHEO e C H 3

x3

4. The ratio kT/kp, the true value of the rnigratory abilities of the one p-tolyl and the two phenyl groups of aldehyde IV, should be capable of calculation if the additional ratios k H / k + and k T o l / k ‘ H can be determined experimentally. The ratio k H / k + may be obtained directly from the contributions of paths A to E which are given in paragraph two of this section

P m E

phenyl groups. Further, one would expect that k T o TABLE IT‘ would be greater than k + , particularly since the tolyl RATIOSO F i l R Y L TO HYDROGEN MIGRATIOSFOR X-ARIOUS group of X3 is migrating to a less stable carbonium CARBOSIUM IONS center than the phenyl of XZ. Provided, therefore, k < r / k i i in ReferHzS04 HCOOH ence Carbonium ion that k H and k ‘ H are not greatly different in magniOH tude, i t does not seem unreasonable that the ratio $1 m d / m e is greater than one, even in the event the p7.33 1.44 9 1. PhaCCHPh tolyl and phenyl groups exert their normal migraOH tional tendenciesI2 of about 3 to 1 in favor of p 6.V 1 . 1 5 This tolyl (this would correspond to a kT/kp ratio of 2. P h H & & - a - C € L Ph paper about 1.5, since aldehyde IV contains two phenyls). OH Now if x1, x2 and x3 are the instantaneous, timevariable concentrations of the ions X1, Xzand XS, 3. P h 2 ? C H - a - C H 3 47-56 9 . 3 This (calcd., see paper respectively, then from Chart I dx?(t)/dt

= kpXi(t)

d ~ g ( t ) / d t=

kTXl(t)

- k+Xp(t) - kHXZ(t) - kTolXa(l) - k’HXg(t)

text)

(1) (2)

This value is the reciprocal of lated as indicated in the text. a

kH/k+

= (0.145) calcu-

Since the reactions of Chart I were, in our experiments, allowed to go to completion, equations 1 and 2 are now integrated between the limits zero and infinity. If we replace the values xl(t)dt,

lyzed reaction kT,l/k’a can be approximated by either of the proportions

xz(t)dt and x,(t)dt with the “integration areas” sl, sq and sa, then, since a t t = 0 and a t t = 03, the concentrations of all intermediates are zero

from which values of 47 and 56 are obtained. Thus

Som

kpsi kTSi

- k+sg -

kTolSg

- kHS? 0 - k ‘ ~ S 3= 0

(3)

(4)

k.roi/k’a: kT,l/k’a:

9.3 = 7.33:1.44 9.3 6.9:1.15

(14) Equation 8 was obtained by use of t h e “area theorem” of John Z. Hearon, which is evident in the derivation of equation 3 from equation 2. T h e authors are indebted t o Dr. Hearon for pointing out t h e utility of this theorem in chemical kinetics and f u r t h e derivation of equation 8. It should be noted t h a t no “steady-state” assumptions have been necessary in arriving a t equation 8 .

Sept. 5 , 19%

4333

ALDEHYDE-KETONE AND PINACOL REARRANGEMENTS CHARTI1

*

O H OH I I* PhC-CHPh I

-

*Ph C O C H P h

*

I 0, PhC OH- C H P h

OH

* a!+

-!%

P h C C H Ph

*

*

PhzCCHO

*

*

PhCHCOPh

e CH3

CH3

Ip

*

t

kTOL

OA O A

OA

* ;hzCHEO

I PhC-

; h 2 $ > H e C H 3

I* CHPCh

OA I + * */"tH3 PhC-?HPh -phCOCfHp\

G

*

e

OA * I PhzC

- kH O

C

H

3

e

k ~ / k p(equation 8) is calculated to be either 1.3 or 1.6, corresponding to p-tolyl/phenyl migration ratios of 2.6 and 3.2, respectively. These latter values are normalll and quite in agreement with the qualitative discussion given earlier in this section concerning the mechanism (Chart I) of the aldehydeketone rearrangement. B. The Formic Acid-catalyzed Reactions.Following are the implications of the data of Tables I, I1 and I11 for the formic acid-catalyzed rearrangements of IV, V and VI. 1. The reactions of IV, V and VI in boiling formic acid must be extremely more complicated than the reactions of these same compounds in sulfuric acid. The most significant difference in the radiochemical results for the two cases lies in the fact that the chain-label of aldehyde IV undergoes scrambling in boiling formic acid. This must mean, therefore, that the majority of the scrambling of the chain label of Va takes place after conversion of Va to the conjugate acid of aldehyde IV (that is, to ion XI). This conclusion follows from lines 2, 4 and 5 of Table 11. To accommodate this conclusion the mechanism outlined in Chart 11-in which the group

A may be either formyl or hydrogen16-is proposed. 2. The contributions of the reaction paths of Chart I1 can now be calculated. If ma to mf represent the moles of glycol V which proceed to product through paths A to F, and neglecting the unim(15) Glycol Va, in formic acid, can probably either esterify its secondary hydroxyl group t o give a formate or lose its tertiary hydroxyl t o form ion XO. T h e direct loss of secondary hydroxyl t o form the corresponding carbonium ion is undoubtedly of insignificant importance. If the esterification reaction is slow compared t o the forrilation of ion Xo, then paths A and B ( C h a r t 11) will likewise be of no importance. Ion Xo can rearrange as shown (Chart 11) t o X I and the IV XI[A = HCO-1. latter can esterify by the process XI [A = H I Internal return [S. Winstein and D. Trifan, ibid., 74, 1154 (1952); S . Winstein and K. C. Schreiber, ibtd., 74, 2165, 2171 (1952); S. Winstein and R. Heck, ibid., 74, 5584 (1952) 1 can t h e n occur as shown in C h a r t 11, thus favoring reformation of t h e secondary carbonium ions with attendant aryl migration t o a greater extent t h a n is possible from t h e unesterified glycol itself. This hypothesis is strengthened b y t h e observation, t o be presented later, t h a t a monoformate of triphenylethylene glycol can be isolated from a formic acid solution of t h a t compound. From a formic acid solution of glycol V itself, a mixture of compounds, probably formates, has also been obtained. If t h e esterification reaction competes favorably with tertiary hydroxyl removal, then internal return can occur before formation of ion Xa, and some scrambling could take place through paths A and

--

B.

BEN11.BENJAMIN AND CLAIRJ. COLLINS

4334

Vol. 78

portant contributions of paths A, B and G,16then i t can be shown (see Experimental section for calculations) that: m, = 0.466, md = 0.366, me = 0.120, mf = 0.040. If the mechanism of Chart I1 is correct, then ( m d mf)/me (which equals 3.45) should be the same, within experimental error, as the ratio of ketones VI1 and VI11 obtained from aldehyde IV itself under these same conditions. That the latter value (3.48, from Table I, line 2) is in such close agreement with our calculated ratio, is thus excellent evidence for the validity of the mechanism. 3. It is now important to determine whether the 9-tolyl/phenyl migration ratio, given by 2kTIkP (from equation 8), can be shown to be greater than one. The ratio kHlk+ may be evaluated directly

+

[2]100°

- md

+ me + mi -Wl,

0.466

0.366

= 0.87 + 0.120 + 0.018= -0.466 0.531

It is interesting that from Table 111, representing the formic acid-catalyzed reaction a t 27.6', the ratio kH/k+ a t that temperature may be determined as

[$],,,,.

=

=

0.96

It thus appears that kH/k+ is reasonably constant between these two temperatures. Lye may calculate also from Table I11

It is possible to determine [ k ~ ~ l / k ' ~ ]from l o o ~the data relative to the rearrangement of VIa of Tables I and I1 and from the calculated values of the contributions of the paths of Chart 11. Given in Chart I11 is an over-all mechanism for the rearrangement of glycol VI, based upon the scheme of Chart 11. Those paths whose contributions are less than a very small percentage of the total reaction (see footnote 15) have been omitted. We have already shown that once ion X1 is formed, whether from the aldehyde IV or from the glycol V, it rearranges to VI1 and VI11 in the same ratio when the conditions of the reaction are maintained constant. It is possible, therefore, to distinguish the contribution of path K from that of path E (Chart I I I ) , since we know the ratio (md mf),/mefrom IV or V under these same conditions. Thus, if mj, ma and mi are the mole fractions of the total reaction product proceeding to ketone VI1 through paths J, D and F, and m k and me are the mole fractions of total reaction product proceeding to ketone VI11 through paths K and E, then (see Experimental section for calculations) : mj = 0.040, mk = 0.093, W ' Ld = 0.596, me = 0.193, mf = 0.078 and [ k T d me m f ) / m k = 9.3, in good ~ ' H ] I O = O ~ (md = 11. agreement with the value [h~oI/k'H]27.6~

+

+

+

(16) Although paths A and B undoubtedly contribute t o t h e over-all mechanism, their total participation is probably less t h a n 1%. Similarly, although there is definite evidence t h a t path G contributes t o t h e mechanism (since 470 of t h e ketone VI11 which is formed from Va contains its carbon-14 in t h e benzhydryl portion of t h e molecule), t h e extent of participation of path G must be only (0.04) X (0.12) or less t h a n 0.5% of t h e total reaction. If one assumes t h a t t h e ratio m d m g is about 3 (the normal ratio of tolyl/phenyl migratory abilities9, t h e n path H must participate t o t h e extent of less t h a n 1.5% of t h e total reaction. For these reasons, we neglect t h e contributions of ' and H in order t o simplify t h e calculation. paths A, B, 2

--

na @ I* Ph2C C H

0

CH3

PhzCHtO O C H 3 P

OA OA

OA

m

E P h CO$HPh

PII

CH,

4. Equation 8, the exact solution for the mechanism of Chart I, obviously requires modification before i t can be applied to the formic acid-catalyzed reaction, the mechanism for which is given in Chart 11. This is true because path F has been recognized as an important contributor to the over-all reaction. It might be anticipated that path F should affect the final ratio kT/kP in equation 8 by only 9%, the extent to which this path contributes to the reaction. Since the products both from paths E and F proceed through ion X3, however, the error involved is much greater than 970 because mr/(me mf) X 100 = 29%. A further difficulty in applying equation 8 to the present data is that the ratio kH/k+ (Chart 11) might not be the same in ions Xo and Xz when A = H and A = HCO. A rough estimate of kT/kP from equation 8 is 2.2, however, using kH/k+ = 0.87, kTolik'H = 9.3 and the product ratio (m, mi)/md = 0.46 This corresponds to a p-tolyllphenyl migration ratio of 4.4. I t seems likely, despite the uncertainties involved, that the true 9-tolyllphenyl migration ratio is thus greater than one. C. General Conclusions.-From the discussion in Parts A and B of this section i t has been shown that whether the conjugate acid of aldehyde IV (ion XI) is generated from I V itself or from glycol V, i t yields the same ratio of ketones VII/VIII. These ratios observed from IV, and calculated from the radiochemical data for the reactions of Va and Vc are, respectively, 4.7 us. 4.8 (sulfuric acid) and 3.3 os. 3.5 (formic acid). The agreement in each case is excellent and offers compelling evidence for the reliability of the mechanisms given in Charts I and 11. The 9-toly1,lphenyl migration ratios in the aldehyde-ketone rearrangement of IV have been estimated from equation 8 as greater than one when either sulfuric acid or formic acid is the catalyst. ,4

+

+

Sept. 5 , 1956

ALDEHYDE-KETOXE AND PINACOL REARRANGEMENTS

4335

ratio of greater than one is normal12for p-tolyl ver-

has been the assumption that the two diastereomers glycol of ring-labeled 1,2-diphenyl-l-p-tolylethylene hyde-ketone rearrangement-which has previously (Vc) proceed to products VI1 and VI11 with identibeen considered anomalous because of an apparent cal consequences. This assumption is not inconr e ~ e r s a ofl ~the ~ ~usual ~ ~ mobility sequence-can be sistent with the work of Curtin (cf. references 7 and explained as another example of an acid-catalyzed 12 and previous papers) on the cis-effect, since even rearrangement in which the normal order of migra- if the reaction Xo-+ X1 (Charts I , I1 and 111) were tory abilities is maintained. This fact is taken as a concerted process, the transition states for this additional strong evidence for the validity of the reaction (corresponding to the two diastereomers of Vc) would be expected to be sterically equivalent, mechanisms of Charts I and 11. The values of k ~ ~ (see / kTable ~ IV) vary with owing supposedly to the equivalent effective bulk of the catalyst employed and are greater for the sul- the phenyl and p-tolyl groups. To the extent also furic acid than for the formic acid-catalyzed reac- that rearrangement of the glycol V approaches the tions. This observation is consistent with those re- limiting carbonium ion mechanism illustrated in ported in a previous paperg and tentatively ration- Charts I, I1 and 111, the cis-effect would be exalized on the basis of a steric repulsion to aryl mi- pected to approach zero. This assumption that gration during the concerted migration of aryl to the diastereomeric glycols Vc rearrange identically form a triarylmethyl group. finds further support (a) in the experimental obFrom the data of Table I1 i t can be seen that the servations (Table I) that the a- and 0-forms (Vb glycol VIa has undergone 28y0 scrambling of its and Va) of the chain-labeled glycol rearrange to VI1 label on conversion, in sulfuric acid, t o VII. This and VI11 with experimentally indistinguishable concorresponds to 23% secondary hydroxyl removal, sequences and (b) in the agreement between the obsince the yield of VI1 in this experiment was 83y0. served ratios of VI1 t o VI11 (4.7 in sulfuric acid and We have concluded (see Part A of the Discussion) 3.5 in formic acid) upon rearrangement of aldehyde that both secondary and tertiary hydroxyl removal IV and the calculated ratios (4.8 in sulfuric acid and from V and VI are irreversible processes in cold, 3.5 in formic acid) obtained from the double-labeling experiments upon glycol V. OH OH Current research upon the pinacol rearrangement Phz-; CH3 P h 2 AI - -? e C H 3 problem is devoted to testing the generality of the mechanism outlined in this paper. H H sus phenyl migration and indicates that the alde-

0 Y

concentrated sulfuric acid; thus the ratio 7 7 / 2 3 = 3.35/1 is a measure of the relative ease of formation, in this solvent, of ions Y and 2. From the data of Table I11 i t is clear that in formic acid a t 27.6' there can be a maximum of 5% secondary hydroxyl removal from VIa (Y/Z = 19/1), since this is the yield under these conditions of VIIb labeled only in 4-methylbenzhydryl portion of the chain. It is clear, therefore, that a tolyl group is much more able to stabilize ion Z relative to Y in concentrated sulfuric than in formic acid. It may be that formic acid, with a weaker ionizing ability than sulfuric, is capable of removing chiefly only the less firmly bound tertiary hydroxyl group of VI. Sulfuric acid, with a high ionizing power, is able t o effect removal of both hydroxyls, being assisted in this by the tolyl group, as symbolized by either of the following;

H

Experimental

2

H

The tolyl group, being a better electron donor than phenyl, might be expected, in sulfuric acid, to have those electrons proportionately more available than those of phenyl to distribute the charge of the positive carbonium center. This hypothesis is supported by the observationg that the action of cold, concentrated sulfuric acid upon triphenylethylene glycol causes a negligible proportion of secondary hydroxyl removal. Implicit in our calculations of the contributions of the various reaction paths of Charts I, I1 and I11

1,2-Diphenyl-l-p-tolylethylene-l-C1~ Glycol (Vb)

a-

Form.17-Carbonyl-labeled desoxybenzoin was converted to desyl bromide by the general method of Staedel and Kleinschmidt .I* The carbon disulfide solvent was removed by distillation in an atmosphere of COZ,and without purification the crude desyl bromide was converted to benzoin-C14 by the method of Wardl9; the yield of material of m.p. 125130" was 79y0. Benzoin-C14, 0.12 mole, was added to the Grignard reagent prepared from 0.37 mole of p-bromotoluene and 0.37 mole of magnesium. After being heated a t reflux temperature for 1 hr., the reaction mixture was treated with ice and ammonium chloride solution. The glycol Va was isolated, dissolved in ethanol and the solution was decolorized with Norite. After four crystallizations the glycol had a m.p. of 167-169.5Om and a radioactivity assay of 0.8630 rt 0.0006 mc./mole. 1,Z-Diphenyl-l-p-tolylethylene-2-C~4 Glycol (Va) 8-Form. -Benzyl p-tolyl ketone labeled in the methylene group was prepared by the Friedel-Crafts synthesis from phenylacetyl%C14 chloride and toluene. I t was converted to 4-methyldesyl bromide and thence to 4-methylbenzoin-CI4 by the methods described in the preceding section. The 4-methylbenzoin (0.033 mole) was added to the Grignard reagent prepared from 0.15 mole of bromobenzene and 0.15 mole of magnesium. When the vigorous reaction had stopped, 200 ml. of tetrahydrofuran was added and the ether was removed by distillation. The reaction mixture was heated a t 50" for 18 hr. and the Grignard complex was hydrolyzed with ammonium chloride solution. The crude glycol Va was extracted from the aqueous layer with ether, the ether was removed in a current of air and the glycol was dissolved in ethanol. After clarification with Norite and four crystallizations from alcohol, the product had a m.p. of 18718SoZ1and a radioactivity assay of 2.286 rt 0.008 mc./mole. (17) We are indebted to Dr. H. J. Schaeffer for preparing this compound. (18) W. Staedel and F. Kleinschmidt, Bey., 13, 837 (1880). (19) A. M.Ward, J . Chem. Soc., 1549 (1929). (20) Reference 10a gives a m.p. 168O; reference 10b gives m.p. 161162O. (21) A. McKenzie and A. L. Kelman, J . Chcm. Soc., 412 (1934), gives a m.p. 181-182O.

4336

BEN XI. BENJAMINA N D CLAIRJ. COLLINS

1,2-Diphenyl-l-p-tolylethyleneGlycol @-Form (Vc), 1Phenyl-labeled.-4-Methylbenzoin (0.018 mole) was added to the Grignard reagent made from 0.028 mole of carbon-14 labeled bromobenzene and 0.03 mole of magnesium. T h c reaction was completed as described for the chain-labeled glycols Va and Vb. T h e phenyl-labeled glycol S'c had a m.p. of 187-188" and a radioactivity assay of 1.209 =t 0.001 mc./mole. 1, 1-Diphenyl-2-p-tolylethylene-2-CL4 Glycol (VI) .Zz-pMethylacetophenone (0.56 mole) labeled in the carbonyl position was oxidized t o p-methylphenylglyoxal with SeO? by a n adaptation of the method of Riley and Gray.13 After removing the selenium by filtration, the p-methylphenylglyoxal was not isolated but was rearranged to pmethylmandelic acid by adding to its ethanol solution 0.56 mole of NaOH in 110 ml. of ethanol. The reaction mixture was stirred for 1hr., then concentrated to 100 ml. by evaporating the ethanol in a current of air. Pure p-methylmandelic acid, m.p. 148', was recovered in 24',; yield. The methyl ester (0.132 mole) of p-methylmandelic acid, prepared from the acid and diazomethane, was added to 0.53 mole of phenylmagnesium bromide. The reaction was forced t o completion in the tetrahydrofuran solution as described in previous sections. T h e glycol VI was purified by five crystallizations from ethanol, m.p. 195-196'. The radioactivity assay was 3.812 i 0.010 mc./mole. Diphenyl-p-tolylacetaldehyde-l-C14 (IV).-1,l-Diphenyl2-p-tolylethylene-2-C14 glycol (VIa), 2 g., was mixed with 300 ml. of 98% formic acid. The mixture was stirred a t room temperature for 24 hr. During this time the solution hecame yellow and the glycol slowly dissolved. \Vater was added until the volume was 800 ml. and the precipitate was extracted twice with ether. The ether solution was extracted with water followed h y sodium bicarbonate solution. The oily material remaining after evaporation of the ether was dissolved in aqueous alcohol. Crystallization was accomplished by allowing this alcoholic solution to stand and slowly evaporate for one week. Several batches of aldehyde prepared in this way were combined and recrystallized twice from aqueous alcohol. The aldehyde had a n1.p. of 60-61" and a radioactivity assay of 3.798 mc./mole. Anal. Calcd. for C21H180:C, 88.08; H, 6.33. Found: C, 88.05; H, 5.92. The 2,4-dinitropheuylhydrazone of aldehyde IS' had a m.p. of 181-183". Anal. Calcd. for C27H2rX10: C, 69.51; H, 4.75. Found: C, 69.40; H, 5.11. Radiochemical Structure Determination of IV, Va, Vb, Vc and VI.-Glycol Vb was oxidized with excess I(MnO4 in acetone. The benzoic acid fraction was discarded and the neutral fraction containing 4-methylbenzophenone was treated with 2,4-dinitrophenylhydrazine reagent. 5 f t e r two crystallizations of the 2,4-dinitrophenyIhydrazone, the assay was 0.7558 3z 0.0001 mc./mole. Thus the tertiary carbon position contained 87.5% of the radioactivity. This distribution of radioactivity was taken into consideration in the rearrangement experiments. Glycol 1-a was oxidized to benzoylbenzoic acid with chromic acid. After purification the benzoylbenzoic acid was assayed and found to be non-radioactive. -411 the radioactivity was in the 2-position. Glycol Vc was oxidized in the same fashion. The assay of the purified benzoylbenzoic acid was found to be 1.204 i 0.001 mc./mole. Glycol VI was oxidized to benzophenone and terephthalic acid with chromic acid. The 2,4-dinitrophenylhydrazonederivative of the benzophenone was found to be non-radioactive. The assay of the dimethyl ester of the terephthalic acid was 3.787 i 0.012 mc./mole. Thus the 2-position contained all the radioactivity in glycol VI. The radiochemical structure of aldehyde IV follows from the sulfuric acid-catalyzed rearrangement t o 1'11 and VI11 and the ensuing degradation of these ketones (Table 11). Rearrangement of Compounds IV, Va, Vb, Vc and VI with Acid Catalysts. ( 1 ) Concentrated Sulfuric Acid.-In a typical experiment 3.116 g. of glycol Va was added to concentrated sulfuric acid, which had been cooled previously to O D . The mixture was stirred by means of a magnetic stirring bar for 2 hr., during which time the reaction flask waq immersed (22) J. F. Eastham, private communication. (23) A. H. Blatt, "Organic Syntheses," Cull. Vul. I I , J o h n Wile?. and Sons, Inc., New York, S. Y . , 1043.p. 20'3.

Val. '78

in a n ice-bath. The contents of the flask were poured over 100 g. of cracked ice, and the resulting slurry was extracted with six 50-ml. portions of ether. T h e ether solution w a i washed with aqueous bicarbonate and then was evaporated to d r y n e s . T h e concentrate was taken up in ethanol and diluted to 250 ml. in d volunietric flask. After the solution was mixed completely, it was divided into three aliquots to be used for determining the yields of VI1 and VI11 and for determining the carbon-I4 distribution in these products. ( 2 ) Boiling Formic Acid.-Glycol S'a (1.040 g.) was mixed with 10 ml. of 98-10070 formic acid and the mixture was heated a t reflux temperature for 20 hr. T h e products were not separated from the acid. The solution waq diluted to 250 ml. in a volumetric flask. ( 3 ) Formic Acid a t 27.6°.-Glycol T'a (2.000 9.) was added t o a flask containing 300 ml. of 98-10070 formic acid which had been in a constant temperature bath a t 27.6' for 1 hr. T h e reaction mixture was stirred a t 27.6' with a magnetic bar for 24 hr. d t the end of this time the clear yellow solution was diluted with twice its volume of water, and the precipitated material was removed by six extractions with ether. T h e ether solution was washed with SOdium carbonate solution and the ether was removed by evaporation in a current of air. There remained a n oily materia1 which was either dissolved in a small quantity of alcohol for determining the yield of aldehyde I V or it was dissolved in a measured volume of alcohol as required for isotopic dilution technique. Yield Determinations. (1) Ketones VI1 and VII1.-An aliquot of the ethanol solut'ion of products from the various acid-catalyzed reactions was added t o a carefully weighed portion of the appropriate non-radioactive ketone. Xfter ensuring complete solution, the solvent was evaporated and the ketone was reisolated and crystallized from ethanol (five times, once after Sorite treatment) until the melting point and radioactivity assays were constant. The yields of V I 1 and VI11 were calculated from the assay data as clescribed previously9 and the results are given in Table I . The experiments are summarized in Table Y. (2) Aldehyde 1V.-To the alcohol solution of the products of a reaction of glycol in formic acid a t 27.6' was added an excess of 2,4-dinitrophenylhydrazine reagent. The mixture !$-as allowed to stand for 2 days a t room temperature. The crystalline 2,4-dinitrophenylhydrazoneof the aldehyde which formed was collected on a weighed sintered glass filter, washed with a small portion of fresh alcohol, dried in vacuum and weighed. The yield of aldehyde was calculated from the weight of this derivative. A sample of the pure aldehyde gave a 9974 yield of its 2,4-dinitrophenylhydrazone when no ketones VI1 and VI11 were present. These ketones do not yield 2,4-dinitrophenylhydrazoneunder these conditions. Determination of Distribution of Carbon-14 in Ketones VI1 and VII1.-In a typical experiment, the solvent was evaporated from an aliquot of the products of a rearrangement reaction. T o the residue was added 25 ml. of 35 potassium hydroxide in methanol. The solution was heat at reflux temperature for 24 hr. IVater was added to the reaction mixture and the neutral organic fraction consisting of diphenylmethane and phenyl-9-tolylmethane was extracted with ether. After evaporation of the ether, tlle residue was dissolved in 10 nil. of acetic acid and to it w a s added a solution of chromic acid in acetic acid. The mixture mas heated on a stearn-bath for 1 hr., after which tiiiie it was diluted with water and the precipitated organic material Tyas extracted with ether, T h e ether solution W:I\ washed with three small portions of sodium carbonate d 1 1 tion and saved. The combined sodium carbonate extracts were treated with Sorite and neutralized with hydrochloric acid. T h e precipitated benzoylbenzoic acid was collected on a filter and recrystallized from aqueous alcohol, dried and assayed for radioactivity. -All of these p-benzoylbcnzoic acid fractions possessed sharp melting points of 196197', when observed on a Kofler hot-stage. The experiments are summarized in Table V I . The ether solutic~n from above was concentrated to recover benzoplieriont.. This was converted to its 2,4-dinitrophenylhydrazone. I n many cases five crl-stallizations were not sufficient to raise the m.p. above 210'. For this reason assal-s for the benzophenone fractions could not always be obtained and thus :ire not included in all of the experiments of Tcihles 11, 111 and 1'1.

4337

ALDEHYDE-KETONE A N D PINACOL REARRANGEMENTS

Sept. ,i I, R X

TABLE V

SUMMARY OF YIELDDETERMINATION EXPERIMENTS Starting compound Va, @-form Va, @-form Vb, a-form Va, @-form Vb, a-form VIa VIa VIa IV IV Va, @-form Va, 0-form

Radioactivity, mc./mole 2 . 2 8 6 10,008 2.243 f ,036 0,8630 & ,0006 2.270 1 ,031 0,8848 f ,0053 3.812 i ,010 3.812 1 ,010 3.812 i ,010 3.800 j= ,013 3.800 1 ,018 2.031 1 ,010 2.031 f ,010

Wt., g. 3.1163 0.6085 0.6185 1.0398 1.1075 2.9987 1.1218 0.8407 1.0988 0.8767 1.0742 0.8715

VIa

3.812

1.0160

1 ,010

Aliquot used to determine yield Experimental Time, of ketone hr. VI1 VI11 conditions 40 ml. of 96% HzSOa a t 0' 4 1/10 1/5 3 . 5 1/4 3/4 20 ml. of 96% HzSOi a t 0' 1/4 3/4 20 ml. of90% HnSOa a t O o 2 10 ml. boiling formic acid 21 1/8 3/8 1/8 3/8 40 ml. boiling formic acid 22 1 l/5 1/5 50 ml. of 96% HzSOi a t O o 30 ml. boiling formic acid 2 1/4 1/4 1/4 3/4 30 ml. boiling formic acid 2 2 1/5 l/5 40 ml. of 96Yc HzSOa a t 0" 20 ml. boiling formic acid 3 l/5 1/5 150 ml. formic acid a t 27.6" 24 1 140 ml. formic acid a t 27.6' for 24 hr. then boil 2 hr. 2/5 3/5 1 150 ml. formic acid a t 27.6' 24

Weight of dead ketone added Radioactivity of diluted ketone t o aliquot VI1 VI11 VI1 VI11 0,5411 1.2506 0,6950 10.0051 0.1572 1 0 . 0 0 1 6 . 1 2 4 6 z t ,0004 ,4841 0.9768 ,4392 f ,0028 ,1523 i ,0002 ,0543 1 ,0002 ,5706 ,9580 ,3223 1 ,0023 ,0631 1 ,0006 ,6545 1 ,5906 , 0 4 0 0 1 ,0003 ,5115 0.9587 ,1616 1 ,0001 ,301 =k ,001 1.2480 1.1280 1.046 f ,002 ,3742 3~ ,0007 1.4656 1.2012 0.7052 f ,0024 ,3092 1 ,0013 ,2859 f ,0001 1.7409 1.9556 , 1 3 1 0 1 ,0006 1.0311 1.0181 ,5390 f ,0010 0.8246 0,8573 ,5505 i ,0001 , 1 6 9 7 3 ~ ,0006 ,6153 i ,0003 1.1857 ,4182 i .0018 0.1728 i. ,0017

1.1256 1.0700 1.0177

,1050 f .0002

TABLE VI

SUMMARY OF RADIOACTIVITY DISTRIBUTION DETERMINATIONS Starting compound

Radioactivity of starting compound, mc./mole

Radioactivity a t tertiary carbon osition, mc.Pmole

Radioactivity of benzoylbenzoic acid fraction of products

Experimental conditions

Radioactivity of benzophenone fraction of products

96% H Z S Oa~t 0" 0.0392 iZ 0.0001 Boiling formic acid ,4377 =t ,0001 0.001 96y0 H&O, a t 0" ,0419 f ,0004 0.0336 f 0.0001 ,702 zk ,0034 Boiling formic acid ,6710 .6744 f ,0042 ,001 Boiling formic acid ,0913 ,0799 1.204 f ,001 96% H z S O ~ at 0" ,6805 f ,0001 1.204 f ,001 Boiling formic acid 0.925 iZ ,001 0.0121 967, H2S04a t 0" 1.059 f .004 .0121 Boiling formic acid 0.6637 f ,0014 ,0121 Boiling formic acid ,626 f ,001 ,0121 Boiling formic acid ,625 f .001 ,0121 Formic acid at 27.6" for 24 hr. .I58 =t .002a The benzoylbenzoic acid was obtained by degradation of the P-tolyldesoxybenzoin from the isotope dilution experiment in the last line of Table V.

IV IV Va, P form Vb, cy form Va, Pform Vc, B form Vc, P form VIa VIa VIa VIa VIa

3.800 f 0.013 3.800 f ,013 2.286 iZ ,008 0.8848 f ,0053 2.270 f ,031 1.209 f ,001 1.209 f .001 3.812 iZ .010 3.812 f ,010 3.812 =t ,010 3.812 =t ,010 3.812 =t .010

Calculations of the Contributions of Each Path of Charts I , I1 and 111. A. Chart I. The Sulfuric Acid-catalyzed Reaction of Glycol V.-If c is the mole fraction of VI1 formed through path C (Chart I), then (0.98 - c) is the mole fraction of VI1 formed through path D (since 25", of the reaction is assumed from line 3, Table 11, to proceed through paths A and B). I t then follows from Chart I that c

+ (0.98 - c)0.5 = 0.563 (line 10, Table 11)

and c = 0.146, (0.98 - c) = 0.834. Since ketone VII, formed through paths -4, C and D, represents 84.97, of the product and ketone VIII, formed through paths B and E, represents 15.1y0 of the product (see Table I, lines 3, 4 and 5) it readily follows that the mole fractions of reactant converted to the products through each path are: ma = 0.017,mb = 0.003, m, = 0.124,md = 0.708,me = 0.148. B. Chart 11. The Formic Acid-catalyzed Reaction of Glycol V.-If c, d and f are the mole fractions, respectively, of ketone VI1 formed through paths C, D and F (Chart I I ) , then

+

+

0.765 = c 0.5d 0.5f (line 11, Table 11) d + f = l - ~ f = (1 - C)O.ll5 d = (1 - ~)0.885 and c = 0.530, d = 0.416 andf = 0.054, thus m,, md, me and mt, the contributions of paths C, D, E and F, are

mc = 0.530 (0.88) = 0.466 = 0.416 (0.88) = 0.366 me = 0.120 mi = 0.054 (0.88) = 0.048 C. Chart 111. The Formic Acid-catalyzed Reaction of Glycol VI.-If mj, md and mf are the mole fractions of the total reaction product proceeding to ketone VI1 through paths J , D and F, and mk and me are the mole fractions of total reaction product proceeding to ketone VI11 through paths K and E, then md

mj

+ m+d me + mr= 0.286 = 0.714 (lines 9 and 10, Table I) (lines 9 and 10, Table I)

mk

= 0.165 (line 8, Table 11) + mi mf +md

mj

mj

md

+ mi

I&[

= md

(from Chart 11, see Section B

and

above)

f *m

= 3.5 (line 2, Table I)

me from which: mi = 0.040, mk = 0.093, md = 0.596, me = 0.193, mf = 0.078. Analytical Determinations.-The carbon and hydrogen analyses were performed by the Huffman Microanalytical Laboratories, Wheatridge, Colorado. T h e carbon-14 assays were performed in the usual4 manner, OAKRIDGE, TENNESSEE